Re: Permission denied error for initdb
От | Tom Dron |
---|---|
Тема | Re: Permission denied error for initdb |
Дата | |
Msg-id | CADQ-uoaTPmSMiT9dNg6mxDfNe2nO=ZfL-X_pnVUMb3z7AM9uDw@mail.gmail.com обсуждение исходный текст |
Ответ на | Re: Permission denied error for initdb (Tom Lane <tgl@sss.pgh.pa.us>) |
Список | pgsql-novice |
Yeah, it's 9.3.15, its the required version for the main software I want to use, but I can try a later version if this doesn't end up working.
For some reason I can't find sysctl ANYWHERE on this system, but just looking at the values in /proc/sys/kernel I get:
shmmni: 4096
shmmax: 18446744073692774399
shmall: 18446744073692774399
I'm guessing it might be odd that shmmax and shmall are the same size?
On Thu, Nov 24, 2016 at 12:58 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
Tom Dron <thomas.e.dron@gmail.com> writes:
> fbbgsa@web1:~/.local$ initdb -D ~/.local/data
> ...
> creating template1 database in /home/.local/data/base/1 ... FATAL:
> shmat(id=18808835) failed: Permission denied
That's really weird. Usually problems appear at the preceding shmget
call, not shmat. I'd guess that somehow the System V shared memory
facility is turned off on that machine, but surely it'd fail at shmget
if so. Still, it might be informative to do "sysctl -a | grep shm"
and see what the shmmax/shmall/shmmni parameters are set to.
Also, I gather from the printout that this isn't a terribly late-model
Postgres. If you could try 9.4 or later, that might conceivably have
better luck because it requires only a very small SysV segment.
(This error doesn't sound like it's dependent on segment size, but
I'm grasping at straws...)
regards, tom lane
В списке pgsql-novice по дате отправления: